# Is this AI biased?

Hunted Gift uses Google Gemini AI to generate personalized gift recommendations
based on your inputs like recipient details, occasion, budget, and interests.
Like all AI systems, Gemini's training data reflects patterns from the real
world, which can include biases related to gender, age, culture, and price
ranges.

While we work with a leading AI provider, we can't guarantee recommendations are
completely free from bias. If you notice suggestions that seem off-target or
inappropriate, regenerate for new options.


HOW THE AI WORKS

When you chat with Hunted Gift, your preferences are sent to Google's Gemini AI
models. The AI analyzes your inputs and generates gift ideas by drawing on
patterns it learned during training. It doesn't learn from your individual
feedback or adjust recommendations based on previous user reports.

The recommendations you see depend entirely on:

 * What you tell the AI about the recipient

 * The occasion and budget you specify

 * Style images or interests you share

 * Patterns the AI learned from its training data


POTENTIAL BIASES TO WATCH FOR

AI recommendation systems can exhibit several types of bias:

Gender stereotypes: The AI might suggest makeup for women or tools for men based
on historical shopping patterns, even when those don't fit your recipient.

Age assumptions: Gift suggestions may skew toward typical age-related products
rather than individual interests.

Cultural blind spots: Recommendations may favor Western products or miss
culturally specific gift-giving practices.

Price anchoring: Budget ranges might influence quality perceptions or product
categories in predictable ways.

The more specific you are in your chat, the better. Instead of "gift for my
mom," try "gift for my mom who loves hiking and photography, budget $100."


WHAT TO DO ABOUT BIASED RECOMMENDATIONS

If recommendations feel stereotyped or miss the mark:

 * Regenerate: Click the regenerate button to get a fresh set of suggestions

 * Add details: Provide more specific interests, hobbies, or style preferences
   to override generic patterns

 * Browse saved lists: Sometimes exploring your saved gifts reveals better
   alternatives

Currently, there's no built-in reporting mechanism for biased recommendations.
The AI doesn't learn from individual user feedback or adjust future suggestions
based on what you skip or save.


GOOGLE'S RESPONSIBLE AI PRACTICES

Google Gemini is built with responsible AI principles
[https://ai.google/responsibility/principles/] that aim to minimize harmful
biases. However, no AI system is perfect, and biases can still appear in gift
recommendations.

Hunted Gift doesn't use your data to train or fine-tune the AI. Your
conversations stay private and aren't shared with Google for model improvement.

If you see recommendations that are inappropriate, offensive, or clearly biased,
contact support@huntedgift.com so the team can investigate.